DNA binding by a new metallointercalator that contains a proflavine group bearing a hanging chelating unit

Article abstract of DOI:10.1002/chem.200601855

The new bifunctional molecule 3,6-diamine-9-[6,6-bis(2-amino-elhyl)-1,6- diaminohexyl]acridine (D), which is characterised by both an aromatic moiety and a separate metal-complexing polyamine centre, has been synthesised. The characteristics of D and its Zn¹¹ complex ([ZnD]) (protonation and metal-complexing constants, optical properties and self-aggregation phenomena) have been analysed by means of NMR spectroscopy, Potentiometric, spectrophotometric and spectrofluorimetric techniques. The equilibria and kinetics of the binding process of D and [ZnD] to calf thymus DNA have been investigated at I =0.11 M (NaCl) and 298.1 K by using spectroscopic methods and the stopped-flow technique. Static measurements show biphasic behaviour for both D-DNA and [ZnD]-DNA systems; this reveals the occurrence of two different binding processes depending on the polymer-to-dye molar ratio (P/D). The binding mode that occurs at low P/D values is interpreted in terms of external binding with a notable contribution from the polyamine residue. The binding mode at high P/D values corresponds to intercalation of the proflavine residue. Stopped-flow, circular dichroism and supercoiled-DNA unwinding experiments corroborate the proposed mechanism. Molecular-modelling studies support the intercalative process and evidence the influence of NH⁺...O interactions between the protonated acridine nitrogen atom and the oxygen atoms of the polyanion; these interactions play a key role in determining the conformation of DNA adducts.

Full text of DOI:10.1002/chem.200601855

DOI: 10.1002/chem.200601855
DNA Binding by a New Metallointercalator that Contains a Proflavine
Group Bearing a Hanging Chelating Unit
Carla Bazzicalupi,[a] Andrea Bencini,[a] Antonio Bianchi,*[a] Tarita Biver,[b]
Alessia Boggioni,[b] Sara Bonacchi,[a] Andrea Danesi,[a] Claudia Giorgi,[a]
Paola Gratteri,[c] Antonio Marchal Ingraín,[d] Fernando Secco,*[b] Claudia Sissi,[e]
Barbara Valtancoli,[a] and Marcella Venturini[b]
Abstract: The new bifunctional mole-
cule 3,6-diamine-9-[6,6-bis(2-amino-
ACHTREUNGethyl)-1,6-diaminohexyl]acridine (D),
have been investigated at I=0.11m
(NaCl) and 298.1 K by using spectro-
scopic methods and the stopped-flow
technique. Static measurements show
biphasic behaviour for both D–DNA
and [ZnD]–DNA systems; this reveals
the occurrence of two different binding
processes depending on the polymer-
to-dye molar ratio (P/D). The binding
mode that occurs at low P/D values is
interpreted in terms of external binding
with a notable contribution from the
polyamine residue. The binding mode
at high P/D values corresponds to in-
tercalation of the proflavine residue.
Stopped-flow, circular dichroism and
supercoiled-DNA unwinding experi-
ments corroborate the proposed mech-
anism. Molecular-modelling studies
support the intercalative process and
evidence the influence of NH+···O in-
teractions between the protonated acri-
dine nitrogen atom and the oxygen
atoms of the polyanion; these interac-
tions play a key role in determining the
conformation of DNA adducts.
which is characterised by both an aro-
matic moiety and a separate metal-
complexing polyamine centre, has been
synthesised. The characteristics of D
and its ZnII complex ([ZnD]) (protona-
tion and metal-complexing constants,
optical properties and self-aggregation
phenomena) have been analysed by
means of NMR spectroscopy, potentio-
metric, spectrophotometric and spec-
trofluorimetric techniques. The equili-
bria and kinetics of the binding process
of D and [ZnD] to calf thymus DNA
Keywords: DNA · kinetics · molec-
ular modelling · proflavine deriva-
tives · zinc
Introduction
and the properties of which as drugs need to be tested.
These new probes can provide a better understanding of the
activity of many drugs and anticancer agents, and can be of
help in developing new diagnostic tools.
Recent developments in biochemistry concern the design of
new molecules that are able to bind and react with DNA,
[d] Prof. A. M. Ingraín
[a] Dr. C. Bazzicalupi, Prof. A. Bencini, Prof. A. Bianchi, S. Bonacchi,
Dr. A. Danesi, Dr. C. Giorgi, Prof. B. Valtancoli
Dipartimento di Chimica, Università di Firenze
Via della Lastruccia 3, 50019, Sesto Fiorentino (Italy)
Fax : (+39)055-4573-364
Departamento de Química Inorgµnica y Orgµnica
Universidad de JaØn
Campus Universitario “Las Lagunillas”, 23071-JaØn (Spain)
[e] Dr. C. Sissi
E-mail: antonio.bianchi@unifi.it
Dipartimento di Scienze Farmaceutiche
Università di Padova
Via Marzolo 5, 35100 Padova (Italy)
[b] Dr. T. Biver, A. Boggioni, Prof. F. Secco, Prof. M. Venturini
Dipartimento di Chimica e Chimica Industriale
Università di Pisa
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Via Risorgimento 35, 56126, Pisa (Italy)
Fax : (+39)050-2219-260
E-mail: ferdi@dcci.unipi.it
[c] Prof. P. Gratteri
Dipartimento di Scienze Farmaceutiche
Università di Firenze
Via Ugo Schiff 6, 50019Sesto Fiorentino (Italy)
184
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 184 – 196

FULL PAPER
Actually, the activity of many anticancer, antimalarial and
antibacterical agents, and that of aromatic carcinogens finds
its origin in intercalative interactions.[1–3] Acridines are a
special class of compounds, not only because of their wide
use in the pharmaceutical and dye industries, but also due to
their interesting chemical and physical properties.[4,5] They
were the first chromophores the noncovalent interactions of
which with DNA were extensively studied by exploiting
their intense fluorescence, and their binding to double
strands is almost universally interpreted according to inter-
calation.[6–9] Numerous studies have shown that the interca-
lative process is much more complex than was previously
thought, and it depends on the intercalating moleculeꢁs
structure.[10] In this context, bifunctional molecules that bear
both an aromatic moiety and a separate metal-complexing
centre appear to be of particular interest.[11–13] The aromatic
residue provides an anchor for the molecule on the polymer
chain by intercalation, whereas the metal centre might inter-
calate as well, or it might interact with the polymeric back-
bone thereby exerting its function.[14] Barton and co-work-
ers[15] have synthesised special metallointercalators with di-
verse functions, from luminescent probes for DNA[16] to
structural probes for RNA.[17] Other complexes have been
designed to be able to recognise specific sites on DNA as
base pair mismatches.[18] Moreover, metallointercalators that
are able to promote the cleavage of the phosphodiester
bond at a selected point of the chain length have been pre-
pared. Platinum(II) complexes that have been appended to
an aromatic residue are other special metal intercalators.
Their possible anticancer activity has made them attractive
synthetic targets.[10,19,20] The intercalative binding of these
complexes to nucleic acids constitute a prerequisite for the
consequent slow attack of PtII on the base nitrogen
atom.[21,22] On the basis of the above-mentioned arguments,
we found it interesting to synthesise a new fluorescent met-
allointercalator and to investigate its interaction with
double-stranded calf thymus DNA. This molecule, 3,6-di-
with DNA through the formation of electrostatic and hydro-
gen-bond interactions with oxygen atoms of the polyanion.
Accordingly, the metal-complex-forming ability of D and
the thermodynamic and kinetic features of the interaction of
D and its ZnII complex with DNA have been investigated.
Results
Ligand protonation and ZnII complexation: It is known that
acridine derivatives give rise to self-aggregation in aqueous
solution with the formation of dimers, or larger oligomeric
species at higher concentrations.[21,23–28] Because the aggrega-
tion process can interfere with all other equilibria, the abili-
ty of D and its ZnII complex to form aggregates was investi-
gated first.
1
The H NMR spectra in D2O that were recorded at differ-
ent pH values and with D concentrations greater than 1”
10ꢀ3 m showed that in alkaline solutions (pH>8.5), all reso-
nances are split (Figure 1). This second set of signals reflects
the formation of a second equilibrium species that has the
same symmetry as the parent one, and the interconversion
1
of which is slow on the NMR timescale. H NMR measure-
ments performed with different ligand concentrations (0.01–
0.001m) and at different temperatures (298–328 K) showed
that the abundance of the new species increases with in-
creasing ligand concentration and decreases with increasing
temperature, while an opposite trend was found for the
other species. These data are in agreement with a dimeric
new species that is in equilibrium with monomeric D. Mole-
cules of D associate with one another and form p-stacking
interactions between the heteroaromatic units; this is denot-
ed by the significant upfield shift that is experienced by the
acridine
protons
(dH1ꢀdH1’ =0.096 ppm,
dH2ꢀdH2’
=
0.046 ppm, dH3ꢀdH3’ =0.042 ppm, pH 10, Figure 1). Further-
more the interconversion between the two species is reversi-
ble with respect to changes in pH, concentration and tem-
perature.
ACHTREUNGamine-9-[6,6-bis(2-aminoethyl)-1,6-diaminohexyl]acridine,
1
from now on denoted as D (dye) is characterised by a pro-
flavine unit with a diethylenetriamine moiety that is able to
chelate metal ions. A novel structural aspect of D is that the
metal-binding site is linked to the C9position of the acri-
dine group instead of the N10 position, as is the case in simi-
lar metallointercalators.[21] Because the N10 atom of D is ex-
pected to be protonated at physiological pH in both the
metal-free ligand and its metal complexes, we found it inter-
esting to analyse the effect of a positively charged N10 am-
monium group on the DNA-binding process. Actually, this
ammonium group might be able to influence the interaction
To support the H NMR information, we recorded UV/vis
absorption spectra of D in aqueous solution (pH 10.5) in the
concentration range of 1”10ꢀ6–1”10ꢀ4 m; this showed that
no self-association effects were observed under similar dilu-
tion conditions.
After taking these results into account, the proton-trans-
fer reactions in aqueous solution were studied with a low
dye concentration of [D]=5”10ꢀ4 m to avoid the formation
of aggregates. Actually, attempts to determine the stoichi-
ometry and stability constants of such aggregates were made
by working at higher ligand concentrations, but no results
were obtained at an acceptable level of confidence. The pro-
tonation constants that were determined by means of poten-
tiometric titrations are reported in Table 1. In the pH range
investigated (pH 2.5–10.5), D behaves as a tetraprotic base,
in spite of the high number of nitrogen atoms that could, in
principle, undergo protonation.
To establish which nitrogen atom is protonated in each
1
protonation stage, we recorded the H NMR spectra of D at
Chem. Eur. J. 2008, 14, 184 – 196
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
185

A. Bianchi, F. Secco et al.
Prior to complexation stud-
ies, the possible self-association
of complex species was investi-
gated by 1H NMR measure-
ments in solutions of ZnII and
D in a 1:1 molar ratio. Al-
though in this case due to
broadening of the signals of the
aliphatic protons that was
caused by coordination of ZnII
to the diethylentriamine moiety
the spectra were not well re-
solved, a splitting of aromatic
1H signals with the formation
of a new set of upfield-shifted
resonances (Supporting Infor-
mation, Figure S3), similar to
that observed for the metal-
free ligand, revealed the pres-
ence of a new complex species
at pH values greater than 6,
and at a complex concentra-
tions of 5”10ꢀ4 m or higher.
The abundance of this species
in solution increases with com-
plex concentration and decreas-
Figure 1. 1H NMR spectra of D (1”10ꢀ2 m) at pH 4 and pH 10. Primed labels refer to dimeric species.
Table 1. Logarithms of ligand D protonation and ZnII complexation con-
es with increasing temperature according to an equilibrium
between monomeric and p-stacked dimeric species.
stants determined in 0.1m NMe4Cl at 298.1ꢁ0.1 K.
logK
10.21(2)[a]
9.20(2)
8.21(1)
2.81(1)
Speciation of all complexes was performed by means of
potentiometric titrations, and it confirmed the dimeric
nature of such species. Table 1 lists the equilibrium constants
of all complexes. As shown by the distribution diagram re-
ported in Figure 2, which was calculated for millimolar con-
centrations of ZnII and D, the two dimeric complexes
([Zn2D2]4+, [Zn2HD2]5+) are minor species under these ex-
perimental conditions, while the [ZnD]2+ complex and its
D+H+Ð[HD]+
[HD]+ +H+Ð
ACHTREUNG
A
U
G
Ð
N
13.59(8)
13.94 (6)
3.70 (9)
N
ACHTREUNG
G
ACHTREUNG
protonated
and
hydroxylated
forms
([ZnH2D]4+
,
2
A
A
Ð
G
2.60 (7)
7.77 (9)
[ZnD(OH)]+) prevail over all of the pH range. The dimeric
species disappear at complex concentrations lower than
1”10ꢀ4 m, and they become predominant as the concentra-
tion is increased. The absence of aggregates under similar
dilution conditions was confirmed by the UV/vis absorption
spectra of D, which was recorded in the concentration range
1”10ꢀ6–1”10ꢀ4 m.
ACHTREUNG
[a] Values in parentheses are standard deviations on the last significant
figure
different pH values. The pH dependence of the 1H NMR
signals (Supporting Information, Figure S1) showed that the
first two protonation steps (logK=10.21, 9.20) take place on
the primary nitrogen atoms of the diethylenetriamine group,
while the third (logK=8.21) and the fourth (logK=2.81)
protonations involve the acridine nitrogen atom and the ter-
tiary nitrogen atom of the diethylenetriamine group, respec-
tively, according to the Scheme 1. Such protonation behav-
iour is in agreement with the greater basicity of aliphatic
amines compared to aromatic ones, and with the general be-
haviour of aminoacridine.[29,30]
DNA-binding equilibria: On the basis of the previous re-
sults, appropriate concentrations of dyes (1”10ꢀ5 m) and pH
(pH 7 for D, pH 8 for [ZnD]) were selected to ensure the
absence of dye aggregates in the solutions that were used to
study the interaction of D and its ZnII complex with DNA.
The interaction is revealed by changes in the fluorescence
spectra that occur when increasing amounts of DNA are
added to a dye solution.
The involvement of the acridine nitrogen atom in the
third protonation step was confirmed by the absorption and
fluorescence emission spectra that were recorded at differ-
ent pH values (Supporting Information, Figure S2).
A fluorescence emission study was performed with the
metal-free ligand at pH 7, in which [H3D]3+ is almost the
unique species in solution. This showed a sharp decrease of
the emission intensity upon the early additions of DNA to
186
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 184 – 196

DNA Binding by a Metallointercalator
FULL PAPER
From now on, we shall indi-
cate the first branch of the titra-
tion curve (decrease of fluores-
cence) as binding mode 1 and
the second branch of the curve
(enhancement of fluorescence)
as binding mode 2.
The binding process can be
represented by the apparent re-
action given in Equation (1) in
which the dye (D) interacts
with the DNA (P) to give the
bound species (PD). The equi-
librium constant of this reaction
is defined as K=[PD]/([P]”
[D]).
Scheme 1. Protonation pattern of D.
K
PþD PD
ð1Þ
G
H
Figure 2. Distribution diagram of the ZnII complexes that are formed by
D as a function of pH (CZn(II) =CD =1”10ꢀ3 m). I=0.10m (NMe4Cl),
298.1 K.
D (phase 1) followed by an intensity increase (phase 2) and
a shift of the emission maximum towards smaller wave-
lengths (Figure 3, top). Figure 3 (bottom) shows the fluores-
cence titration performed at I=0.11m, pH 7, lexc =385 nm,
lem =500 nm and T=298.1 K. The binding isotherm eviden-
ces the biphasic behaviour of the system, and reveals the oc-
currence of two different modes of interaction.
The interaction of DNA with the ZnII complex of D has
been investigated as well. The measurements were per-
formed at pH 8 at which the complex [ZnD]2+ is practically
the only species in solution. The high stability constant of
this species (Table 1) ensures that the complex is completely
formed prior to reaction with DNA and that the metal ion
is retained by D, thus the spectral changes observed after
mixing the preformed complex with a solution of CT-DNA
only reflect the occurrence of a polymer–dye binding reac-
tion.
Figure 3. Spectrofluorimetric analysis of the binding of D to DNA; CD =
1”10ꢀ6 m, CP in the range 0–1.75”10ꢀ4 m, I=0.11m (NaCl), pH 7,
298.1 K, lexc =385 nm. Top: Fluorescence spectra collected during titra-
tion; the spectral changes reveal the biphasic behaviour of the binding
process. Bottom: Binding isotherm recorded at lem =500 nm; the two
phases, 1 and 2, are associated to two different binding modes.
To evaluate the binding parameters of the two binding
modes, the two branches of the titration curves were ana-
lysed according to Equation (2), in which CD and CP are the
total dye and polymer concentrations respectively; DF=
FꢀfDCD is the change of fluorescence (F) during titration;
fD =F8/CD, in which F8 denotes the initial fluorescence of
The spectral behaviour of the [ZnD]–DNA system (Sup-
porting Information, Figure S4) is similar to that of the
D–DNA system and the features of the binding isotherms
evidence biphasic behaviour for this system as well.
Chem. Eur. J. 2008, 14, 184 – 196
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
187

A. Bianchi, F. Secco et al.
the dye solution and Df=fPDꢀfD is the amplitude.
DF=CD ¼ ðKDf  CPÞ=ð1þKCPÞ
ð2Þ
The analysis of the branches of the titration curves allows
us to evaluate K and Df for the two binding modes (K1,
Df1 and K2, Df2 respectively). The values of K1 and K2 are
collected in Table 2. The quality of the fit is shown in
Figure 3 (bottom), for which the continuous line is calculat-
ed by using Equation (2).
Table 2. Reaction parameters for the interaction of CT-DNA with D and
[ZnD], 298.1 K.
Dye
I [m]
10ꢀ6 K1 [mꢀ1
]
10ꢀ4 K2 [mꢀ1
]
0.01
0.11
1.0
1.7
1.3
0.41
24
6.9
<1.0
D
0.01
0.11
1.0
6.920
5.7
4.3
A
5.6
3.7
0.10
0.20
0.24
–
–
–
6.6[31]
2.0[32]
proflavine
4.3[33] (283 K)
Figure 4. Top: Binding isotherms in fluorescence detection for the
~
*
D–DNA system at different ionic strengths. &: I=0.01m; : I=0.11m;
Salt effect: Fluorescence titrations have been carried out at
different salt concentrations. Because D and [ZnD] are posi-
tively charged ([H3D]3+, [ZnD]2+) under the experimental
conditions adopted for DNA binding, a salt dependence of
the equilibria is likely to occur. In effect, the binding iso-
therms of the D–DNA system show a dependence on ionic
strength going from I=0.01m to I=1.01m (NaCl). At I=
0.01m and I=0.11m the biphasic nature of the binding pro-
cess is evident, whereas at I=1.01m a monophasic behav-
iour is observed (Figure 4, top).
I=1.01m. lem =500 nm, pH 7.0, 298.1 K. Bottom: Binding isotherms in
fluorescence detection for the [ZnD]–DNA system at different ionic
~
*
strengths. &: I=0.01m; : I=0.11m;
I=1.01m. lem =500 nm, pH 8.0,
298.1 K. CD =C[ZnD] =1”10ꢀ6 m.
for the [ZnD]–DNA system this ratio is about four-times
smaller.
Absorbance measurements: Absorbance titrations confirm
the biphasic behaviour described above. Absorbance spectra
of the D–DNA system were recorded during titrations, and
showed that the absorbance first decreases (phase 1, which
corresponds to binding mode 1) and then increases (phase 2,
which corresponds to binding mode 2) and displays a batho-
chromic shift (Supporting Information, Figure S5A). A simi-
lar feature is exhibited by the [ZnD]–DNA system (Sup-
porting Information, Figure S5B).
Two modes of binding are also present at I=1.01m
(Figure 4, bottom) for the [ZnD]–DNA system; this reveals
that the metal ion plays a role in the interaction of the zinc
complex with DNA. Table 2 summarises the results.
Dye displacement assay: To better understand the nature of
the binding process, a fluorescent intercalator displacement
assay[34] was applied to the D–DNA and [ZnD]–DNA sys-
tems by adding CT-DNA, which had been previously satu-
rated with the intercalating cyanine dye (Cyan40), to sam-
ples of D or [ZnD]. The cyanine dye strongly interacts with
DNA by intercalation, and in doing so hinders the penetra-
tion of further agents between base pairs. Moreover, its
emission spectrum does not overlap with those of D and
[ZnD]. The titrations show that binding mode 1 is still pres-
ent, while the binding mode 2 is strongly reduced (Figure 5).
Table 3 reports the values of the ratio Df2/Df1 between
the amplitudes of binding mode 2 and binding mode 1 in the
absence and in the presence of the cyanine. For the D–DNA
system, the amplitudine ratio in the presence of cyanine is
ten-times smaller than that found with DNA alone, whereas
Circular dichroism measurements: The CD spectra of
D–DNA and [ZnD]–DNA have been recorded under condi-
tions for which binding mode 2 is prevalent. Also the CD
spectrum of the proflavine–DNA system has been recorded
for comparison (Figure 6). The CD spectra of D–DNA (Fig-
ure 6a) and [ZnD]–DNA (Figure 6b) are quite similar, and
their shapes resemble that of proflavine intercalated into
DNA (Figure 6c) in that the positive band appears at longer
wavelengths with respect to the negative one.[35] The shifts
of spectra a) and b), with respect to c), along the wavelength
axis depends on the position of the absorption band of pro-
flavine, which is less energetic than the absorption bands of
188
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 184 – 196

DNA Binding by a Metallointercalator
FULL PAPER
Unwinding measurements: To obtain further information on
the DNA-binding mode, an unwinding assay was performed.
Incubation of supercoiled DNA with topoisomerase I gives
rise to relaxed DNA, and upon removal of the protein the
relaxed state is maintained. In contrast, when an intercalat-
ing agent is present in the reaction mixture it alters the
topological state of circular-closed DNA, and upon removal
of both dye and enzyme, supercoiled DNA is suddenly re-
generated. The magnitude of superhelicity is determined by
the original amount of bound dye and the unwinding angle
at the time of the enzymatic religation.[36] An example of a
supercoiled-DNA unwinding experiment is reported in
Figure 7. The presence of a supercoiled plasmid is clearly
Figure 7. Topoisomerase I relaxation reactions of supercoiled pBR322
DNA performed in the presence of increasing concentrations of D and
[ZnD]. S refers to supercoiled pBR322 DNA;
pBR322 DNA obtained by incubating the plasmid DNA with the enzyme
in the absence of dye.
C refers to relaxed
~
Figure 5. Top: Fluorescence titration of D in the presence of DNA (
)
*
and DNA previously saturated with Cyan40
(
). I=0.01m, pH 7.0,
298.1 K and lem =500 nm. Bottom: Fluorescence titration of [ZnD] in
~
*
presence of DNA ( ) and DNA previously saturated with Cyan40 ( ).
I=0.01m, pH 8.0, 298.1 K and lem =500 nm. CD =C[ZnD] =1”10ꢀ6 m.
evidenced starting from 1 mm of D or [ZnD]; this confirms
their ability to bind DNA through an intercalation mode. It
is interesting to note that D and [ZnD] showed comparable
efficiency in altering the enzyme-induced relaxation of the
supercoiled plasmid; this suggests comparable stability of
the intercalated complexes.
Table 3. The ratio between the amplitude of the binding mode 2 and the
binding mode 1 in a traditional titration, Df2/Df1, and in presence of
DNA saturated with Cyan40, (Df2/Df1)Cyan40. lem =500 nm, 298.1 K.
Df2/Df1
ACTHERU(GN Df2/Df1)Cyan40
D–DNA
ACHTREUNG[ZnD]–DNA
ꢀ0.34
ꢀ0.50
ꢀ0.029
ꢀ0.13
Kinetics: The kinetics of the interaction between D and
DNA were investigated in a range of dye and polymer con-
centrations at which binding mode 2 is operative. Relaxation
times (t) were measured under pseudo-first-order conditions
(CP ꢂ10CD) and kinetic traces were fitted by mono-expo-
nential functions (Supporting Information, Figure S6). A
plot of the reciprocal time constant (1/t) as a function of
DNA concentration (CP) shows that the kinetic effect is in-
dependent of the polynucleotide concentration (Figure 8,
top).
The [ZnD]–DNA system has been studied under analo-
gous conditions. The analysis of the data reveals that also in
this case the kinetics are independent of the DNA concen-
tration (Figure 8, bottom).
Figure 6. Circular dichroism spectra of a) D–DNA, b) [ZnD]–DNA and
c) proflavine–DNA systems. CD =3.0”10ꢀ5 m, CP =3.0”10ꢀ5 m, pH 7.0, I=
0.11m (NaCl) and 298.1 K.
Molecularmodelling : The conformational analysis of the
ligand molecule was carried out before the in silico investi-
gation of its binding mode to the biological target. In partic-
ular the attention was focused on the [H3D]3+ species, which
is predominant in aqueous solution at the pH value (pH 7)
that was used for the binding studies.
D and [ZnD]. Similarly, the higher amplitude of the profla-
vine band can be ascribed to the higher absorptivity of this
dye. This finding is in agreement with the intercalative fea-
tures of binding mode 2.
Chem. Eur. J. 2008, 14, 184 – 196
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
189

A. Bianchi, F. Secco et al.
non-intercalative binding modes and by investigating the in-
teraction of the ligand with both the minor and major
grooves of DNA. Moreover, a set of five characteristic basis
sequences that featured 60% CG and 40% AT base-pair
composition, analogous to that of CT-DNA, was considered.
By taking into account the different DNA sequences and
the different interacting modes, several docking experiments
of the three [H3D]3+ conformers were planned and per-
formed. Appropriate experimental settings and conditions
were used to avoid effects on the docking results due to
ligand interaction with the polymer ends. Furthermore, both
CG and AT sites were considered for the intercalative bind-
ing mode. The DE energies for the different poses, namely
the predicted adduct structures, which actually give rise to
favourable interaction (Supporting Information, Table S1)
were calculated according to Equation (3).
X
DE ¼ Eadduct
ꢀ
Ereagent
ð3Þ
Inspection of the DE values clearly evidences that the
main binding mode for the considered P/D ratio should be
intercalation. However, two different intercalative modes
should be available, parallel and perpendicular, when the
proflavine group intercalates in a parallel or perpendicular
fashion to the base-pair hydrogen bonds.[40] In the parallel
mode the intercalator interacts with both DNA filaments,
while in the perpendicular mode only one filament is mostly
involved in the p-stacking interaction.
Figure 8. Top: Dependence of the reciprocal relaxation time, 1/t, on the
polynucleotide concentration for the D–DNA system. I=0.11m (NaCl),
298.1 K, pH 7.0, lex=405 nm (laser). Bottom: Dependence of the recipro-
cal relaxation time, 1/t, on the polynucleotide concentration for the
[ZnD]–DNA system. I=0.11m (NaCl), 298.1 K, pH 8.0, lex=405 nm
(laser). CD =C[ZnD] =4”10ꢀ6 m.
The two intercalative binding modes are depicted in
Figure 9, which shows adducts obtained with the
GCGCGCGCGC (left) and CGCGATATCG (right) se-
quences. Selected structural information (Table S2) and fig-
ures (Figure S7) for all calculated adducts are reported in
Supporting Information.
Analysis of the structural parameters of acridine mole-
cules that had been functionalised with Car-NH-Cal groups
(Car and Cal stand for aromatic and aliphatic C atoms, re-
spectively) linked to the C9position was carried out with
the ConQuest[37] and VISTA programs[38] on the Cambridge
Structural Database.[39] This showed mainly sp2 character for
this type of nitrogen atom, although a partial loss of conju-
gation between the nitrogen atom and the aromatic moiety
was commonly found. The most frequent value for the
Car-Car-NH-Cal dihedral angle is about 208, which is usually
correlated with a Car-NH-Cal bond angle ranging from 125–
1358.
As a consequence, a Nsp2 atom type was assigned to this
nitrogen atom, and both dihedral and angular values were
constrained to the above-mentioned values.
Three conformers (Supporting Information), each from
one of the most populated clusters, were selected to study
the binding of [H3D]3+ to the biological target. All the con-
formers had antiperiplanar torsional angles of the C6 chain
(trans conformation) and mainly differed from each other in
the orientation of the diethylentriamine moiety with respect
to the acridine plane, which was almost perpendicular in
two of them, and almost coplanar in the third one.
Figure 9. Adducts obtained with GCGCGCGCGC (left) and CGCGA-
TATCG (right) sequences showing parallel and perpendicular intercala-
tive binding modes, respectively. DE values are calculated interaction en-
ergies.
The modelling study of the interaction between the ligand
and the selected DNA decamers (see Experimental Section)
was carried out by considering both the intercalative and
190
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 184 – 196

DNA Binding by a Metallointercalator
FULL PAPER
It should be noted that in the perpendicular binding
mode, hydrogen-bond and electrostatic interactions are
often present that involve the protonated acridine nitrogen
and oxygen atoms of the polymer. On the other hand, the
polyamine chain, which is accommodated in the major or in
the minor groove of the different adducts, contributes to the
overall stability through the formation of one or two hydro-
gen-bond contacts with the phosphate groups.
to analyse the nature of the aggregate, because, as shown by
1H NMR measurements, the formation of this species takes
place at concentrations greater than 1”10ꢀ3 m of D; this
concentration range corresponds to absorbance values that
are not suitable for binding studies. In the case of the ZnII
complex, although similar problems were found in the anal-
ysis of self-aggregation by means of absorption spectra, po-
tentiometric measurements make it possible to establish the
dimeric nature of the aggregates and determine their forma-
tion constants (Table 1).
Several adducts that are derived from nonintercalative
binding were also predicted by the molecular-modelling cal-
culations. As an example, Figure 10 shows the adduct that
1
The analysis of the H NMR spectra of the aggregates fur-
nishes some indications about the relative orientation of the
adjacent molecules in these aggregates. In the case of the
metal-free ligand, the acridine protons and the proton of the
hydrogen atom H4, which is linked to the carbon atom in
the proximity of the acridine group, are subjected to the
shielding effect of the aromatic p-clouds, as denoted by the
1
significant upfield shift of the corresponding H NMR reso-
nances (Figure 1). The two signals that shift the most are
those of H1 and H4 (dH1ꢀdH1’ =0.096 ppm, dH4ꢀdH4’
=
0.126 ppm, Figure 1), which indicates that these two hydro-
gen atoms penetrate more deeply into the shielding region
of the aromatic groups with respect to H2 and H3
(dH2ꢀdH2’ =0.046 ppm, dH3ꢀdH3’ =0.042 ppm, Figure 1).
A
possible arrangement of the aggregate that would account
for the observed 1H NMR behaviour is depicted in
Figure 11. This shows a D dimer in which two face-to-face
Figure 10. Adduct obtained with the CGCGATATCG sequence featured
by major-groove-binding mode. DE is calculated interaction energy.
was obtained for the CGCGATATCG sequence (Supporting
Information, Tables S1 and S2) in which the protonated
ligand lies in the major groove. The adduct is stabilised by
hydrogen-bond interactions with the phosphate oxygen of
the polyanion that involved both protonated acridine and
Figure 11. Proposed arrangement of D dimers.
aliphatic
G
p-stacked molecules, disposed in a head-to-tail mode, have
the centres of the adjacent aromatic rings laterally displaced
towards H1 and H4. Such an arrangement would cause the
chelating hanging groups to be located in the deshielding
zone of the aromatic systems, which is in agreement with
the downfield shift that is experienced by the relevant
1H NMR signals in the aggregate with respect to the mono-
meric molecule (Figure 1). Similar considerations can be
made for the ZnII complex of D, the aggregation mode of
which is probably quite similar to that of the metal-free
ligand.
Analogous interaction modes were also reported for pro-
flavine,[41] acridine orange[21,42] and acridine orange deriva-
tives.[26,43]
This study has shown that the bifunctional molecule D
and its ZnII complex ([ZnD]) interact with double-stranded
DNA, to form stable complexes. Static measurements clear-
base and the acridine ligand moiety (CH3···C mean dis-
tance=3.9Š).
Self-aggregation of acridine orange,[23] proflavine and other
similar dyes[24] was shown to produce modifications of the
electronic absorption spectra of the monomeric species that
were interpreted by assuming the formation of dimers in sol-
utions at low concentration, as the principal species in equi-
librium with the monomeric forms. Accordingly, it seems
reasonable that the unique oligomeric species that is formed
by D under our experimental conditions is a dimer, which is
in agreement with the H NMR spectroscopy data. Unfortu-
nately, in the case of D, absorption spectra cannot be used
Chem. Eur. J. 2008, 14, 184 – 196
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
191

A. Bianchi, F. Secco et al.
ly show a biphasic behaviour for both systems, which reveals
the occurrence of two different modes of binding.
ducts formed under the conditions of binding mode 1 (low
dye-to-DNA molar ratios).
The decrease of fluorescence emission and absorbance
signals, observed for the first branch of titrations (binding
mode 1), persists varying the ionic strength. The values of
K1 related to this mode of binding are two orders of magni-
tude higher than those expected for intercalation reactions
in spite of the fact that the approximation made in Equa-
tion (2) ([P]=CP) would cause K1 to be underestimated.
One can note that the values of K1 for both D–DNA and
[ZnD]–DNA systems display a slight ionic-strength depend-
ence, possibly due to charge effects.[44] The major difference
between the two systems lies in the fact that the affinity of
[ZnD] for DNA is near four times greater than that of D;
this suggests that, in binding mode 1, the residue that con-
tains the zinc ion plays an important role.
The kinetics of D and [ZnD]-binding to DNA according
to binding mode 2 show that the values of the time constants
of the two systems are similar; this suggests that the rate-de-
termining step involves a residue common to both dyes, that
is, the proflavine residue. This is in agreement with the com-
parable binding constant that is observed in the presence/ab-
sence of the metal ion for the second equilibrium, which
gives rise to comparable efficiency in DNA unwinding. The
values of the rate constants are an order of magnitude lower
than the value expected for diffusion-controlled reactions,[46]
this fact confirms that the binding process, which is repre-
sented by the apparent reaction (1), is not a simple one, but
is composed of at least two steps.
The kinetic features of binding mode 2 could be rational-
ꢀ
The enhancement of fluorescence emission and the ob-
served red shift of the absorbance spectra during the second
part of the titrations (binding mode 2) indicate a base–dye
interaction. The values of the related binding constant, K2,
are close to the value that was measured for the DNA–pro-
flavine system (Table 2), which is known to bind double-
stranded nucleic acids by intercalation.[31–33] Because D and
[ZnD] bear +3 and +2 charges ([H3D]3+, [ZnD]2+]), re-
ised on the basis of the reaction Equation (4), in which P D
is a bound form that is different from the intercalate PDint.
According to Equation (4) the kinetic data could be fitted
to Equation (5)
k1
Kex
PþD
PꢀD
H
PD
H
ð4Þ
ð5Þ
G
G
int
k
ꢀ1
1=t ¼ ðKexk1CPÞ=ð1þKexCPÞþkꢀ1
spectively, the slopes of logK versus ꢀlog
[Na+] plots could
ACHTREUNG
be much higher (ffi3 and ffi2). The lower values that were
obtained suggest that not the entire molecule, but only the
proflavine residue takes part in binding mode 2. Actually,
the charge of this residue is +1 for D and smaller than 1
for [ZnD]. For these reasons we suggest that binding
mode 2 corresponds to the intercalation of the aromatic resi-
due into DNA, whereas the first part of titration (binding
mode 1) is related to the formation of an external complex
between the dye and DNA.
If KexCP @1, it turns out that 1/t=(k1+kꢀ1) and its con-
centration dependence could not be observed.
Structural information on the adducts that were formed
with DNA were obtained by means of modelling studies
that were performed with DNA sequences (decamers),
which are representative of the calf thymus DNA base-pair
composition. A 1:1 decamer/D molar ratio was adopted that
corresponds to the range of concentration at which the bind-
ing mode 2 was observed. By taking into account the similar
behaviour of D and [ZnD] in DNA-binding mode 2, only
modelling studies based on D were performed.
The dye displacement assay shows for both systems that
upon the addition of DNA saturated with an intercalating
dye, the first binding stage still remains whereas the second
one is suppressed (Figure 5). This behaviour indicates that
D is not able to push the intercalated dye out of the nucleo-
tide cavities, otherwise a fluorescence increase would be ob-
served at the wavelength of 500 nm at which only the
D–DNA complex, which is formed according to binding
mode 2, is fluorescent. Because it is demonstrated that
Cyan40 is intercalated between the base pairs of DNA,[45]
the binding mode 2 necessarily should correspond to an in-
tercalative process. This hypothesis was confirmed by the
unwinding assay as well as by CD experiments. Similar con-
siderations can be applied to the [ZnD]–DNA system.
The observation that the first branch of the titration
curves is not influenced by the presence of cyanine interca-
lated into DNA suggests that in binding mode 1 the interac-
tion between the dye and DNA occurs externally to the
DNA cavities, with the polyamine residue possibly lying
along the grooves of the nucleic acids. The calculated model
depicted in Figure 10 could provide a general idea of this
type of non-intercalative interaction, although it is not rep-
resentative of the specific structural features of D–DNA ad-
The proton localisation that was determined for [H3D]3+
by means of H NMR measurements in solution (Scheme 1)
was used in both a preliminary conformational search and in
docking procedures, by assuming that no changes occur
upon interaction with DNA.
1
As previously reported, the nitrogen atom of the
Car NH Cal linkage has mostly a sp2 character, and there-
fore it should have neither hydrogen-bond acceptor proper-
ties nor donor ability towards metal ions. Actually, similar
binding processes are found in both D–DNA and
ZnD–DNA systems, in agreement with the fact that the
ꢀ
ꢀ
ꢀ
ꢀ
Car NH Cal group is not involved in metal coordination.
The analysis of modelling results (see Supporting Infor-
mation, Tables S1 and S2) shows that intercalation is the
preferred binding mode, in agreement with the solution re-
sults obtained for the binding mode 2 with calf thymus
DNA and with literature data on other acridine-type com-
pounds.[47] Intercalative binding involves both AT and CG
base pairs, even if CG sequences seem to be preferred to
AT ones. Most likely, in the case of the polyA–polyT se-
quence, major-groove binding should occur also.
192
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 184 – 196

DNA Binding by a Metallointercalator
FULL PAPER
Inspection of the geometrical features of the calculated
adducts reveals that both parallel and perpendicular stack-
ing modes (Figure 9) should occur in the D–DNA system,
the perpendicular one is favoured by additional NH+···O in-
teractions between the protonated acridine nitrogen atom
and the oxygen atoms of the polyanion.
acridine nitrogen and the oxygen atoms of the polyanion.
Most likely this peculiar binding feature is favoured by the
structure of the intercalator, which was designed with the in-
tention of inserting the chelating hanging group onto the C9
atom of acridine in order to allow the N10 atom to project
towards the polymer chain.
The perpendicular stacking with acridine-type compounds
is rather surprising, because they are usually reported as
parallel base-stacking intercalators.[47] For instance, the acri-
dine-type drug 9-amino-DACA, which interacts with DNA
in its protonated form, behaves as a parallel bis-intercala-
tor.[48] However, it is worth noting that in the case of 9-
amino-DACA an intramolecular hydrogen bond that in-
volves the protonated heteroaromatic nitrogen of the drug
avoids the formation of hydrogen-bond and/or salt-bridge
interactions with the biological target. This is not the case of
[H3D]3+, which forms additional NH+···O interactions in
several D–DNA species, which enhance the stability of per-
pendicular adducts.
Nevertheless, the perpendicular intercalation mode has
been proposed for several intercalators such as doxorubicin
and daunomycin[49] as well as for acridines[50,51] for which it
was denoted as a half-intercalation. According to this
model, the dye intercalates between two adjacent bases of
one of the two strands, while the positive charge interacts
electrostatically with the phosphate groups of the DNA
backbone.[52,53]
These results may help in the design of new intercalators
in support of improved DNA-binding processes.
Experimental Section
General details: Stock solutions of D (2”10ꢀ3 m) were prepared by dis-
solving weighed amounts of the solid in pure DMSO, and were kept in
the dark at 255 K. Working solutions were obtained by diluting the
stocks to such a level that the DMSO content could be neglected, and
were used shortly after preparation. A solid sample of Cyan40[54] was pro-
vided by Prof. S. Yarmoluk and the relevant stock solution was prepared
as described for D.
Calf thymus DNA was purchased from Pharmacia Biotech (Uppsala,
Sweden), in the form of the lyophilised sodium salt, dissolved into water
and sonicated as described below. Stock solutions were standardised
spectrophotometrically, by using e=13200 molꢀ1 dm3 cmꢀ1 at 260 nm, I=
0.1m (NaCl), pH 7 as obtained from the sample certificate. The polynu-
cleotide concentrations are expressed in molarity of base pairs and are
indicated as CP. Sodium chloride was used to adjust the ionic strength,
and sodium cacodylate (1”10ꢀ2 m) was employed to keep the pH of the
solutions to the desired value.
The D–DNA system was studied at pH 7.0 and the [ZnD]–DNA system
at pH 8.0. All solutions were prepared with double-distilled water, which
was also used as the reaction medium.
DNA sonication was carried out by using an MSE-Soniprep sonicator
(Sanyo Gallenkamp, Leicestershire, UK), by applying seven repeated
cycles of 10 s sonication and 20 s pause to a DNA sample (10 mL of CT-
DNA ꢃ2”10ꢀ3 m), at an amplitude of 14 mm. The sonicator tip was intro-
duced directly into the solution; this solution was kept in an ice bath to
minimise thermal effects due to sonication. Agarose gel electrophoresis
tests indicated that the polymer length was reduced to about 800 base-
pairs.
Conclusion
The new proflavine-based ligand that contains a hanging
chelating diethyelentriamine moiety and its ZnII complex
bind to calf thymus DNA according to a biphasic behaviour
that is defined by two different binding modes depending on
the polymer-to-dye molar ratio (P/D); the ligand forms
complexes of considerable stability. At relatively low P/D
values, both the metal-free ligand and its ZnII complex asso-
ciate with DNA externally to the base pairs; the principal
involvement comes from the polyamine moiety. The pres-
ence of the metal ion leads to a fourfold increase of the dye
affinity for the polymer. On the other hand, at high P/D
values the interaction takes place through intercalation of
the proflavine residue between base pairs; significant effects
are not derived from the presence of ZnII.
1H and 13C NMR spectra were recorded at 298 K on a Varian Unity
1
300 MHz instrument. In H NMR spectra peak positions are reported rel-
ative to HOD at d=4.75 ppm (D2O) or to TMS (CD3OD, CDCl3), while
in 13C spectra peak positions are reported relative to 1,4-dioxane at d=
67.2 ppm (D2O) or TMS (CD3OD, CDCl3). In the pH-metric titrations,
small amounts of 0.01m NaOD or DCl were added to adjust the pD. The
pH was calculated from the measured pD value by using the relationship
pH=pDꢀ0.40.[55]
ESI mass spectra were performed with a ESI Mass TSQ 700 FINNIN-
GAM by using the free ligand D, which was obtained by extraction in
chloroform of an aqueous alkaline solution prepared from
D·4HBr·EtOH.
Synthesis of ligand (N9-{6-[bis(2-aminoethyl)amino]hexyl}-acridine-3,6,9-
triamine) (D): Ligand D was obtained as a hydrobromide salt (D·4HBr),
according to the reaction sequence shown in Scheme 2. Compound 5 was
prepared by a reported method.[56]
Molecular-modelling calculations were performed with
appropriate base pair sequences and showed that the inter-
calative binding should be more favourable when the CG
base pairs are involved, even if intercalation between AT
base pairs also leads to stable complexes. Most likely, in the
case of the AT sequence, significant groove binding also
occurs.
In contrast with the common behaviour of base-stacking
intercalators, modelling calculations indicate that the inter-
calative binding of the present dye should mostly involve
only one filament of the double helix (Figure 9, right), prob-
ably due to NH+···O interactions between the protonated
Synthesis of tert-butyl (6-aminohexyl)carbamate (2): Boc2O (31.6 g,
0.145 mol) was added in small portions to a solution of 1,6-diaminohex-
ane (50.5 g, 0.43 mol) in CH2Cl2 (350 mL) at 08C (water/ice bath). The
mixture was maintained overnight at room temperature, then chloroform
(375 mL) was added and the resulting solution was washed twice with a
5% Na2CO3 aqueous solution (750 mL). The organic phase was dried
over anhydrous Na2SO4 and the solvent was removed under reduced
pressure to obtain a sticky residue, which was dissolved in 1m HCl solu-
tion (50 mL) and washed three times with diethyl ether (50 mL). The re-
sulting aqueous solution was alkalinised by the addition of 2m NaOH
Chem. Eur. J. 2008, 14, 184 – 196
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
193

A. Bianchi, F. Secco et al.
7.76 ppm (d, J=8.0 Hz, 4H); 13C
(CDCl3): d=21.6, 26.1, 26.2, 26.6, 40.6,
41.5, 53.0, 127.0, 129.6, 136.7,
143.0 ppm; elemental analysis calcd
(%) for C24H38N4O4S2:
C 56.44, H
7.50, N 10.97; found: C 56.50, H 7.41,
N, 11.04.
Synthesis of N,N’-[9-({6-[bis(2-{[(4-
methylphenyl)sulfonyl]amino}ethyl)-
amino]hexyl}amino)acridine-3,6-diyl]-
dipropanamide (6): A suspension of 4
(1.64 g, 3.2 mmol), 5 (N,N’-(9-chloroa-
cridine-3,6-diyl)dipropanamide)
(1.14 g,
(0.45 mL, 0.0032 mol) in dry aceto-
nitrile (50 mL) was refluxed for 5 h.
0.0032 mol)
and
NEt3
AHCTREUNG
The solvent was then removed under
reduced pressure and the dark-orange
crude compound was subjected to
chromatography on silica gel (CH2Cl2/
MeOH 100:15) mixture. The eluted
fractions were combined and evapo-
rated under reduced pressure. The re-
sulting solid orange compound was
further dried in vacuum at room tem-
perature. Yield: 2.15 g, 81%; 1H NMR
(CD3OD): d=1.23 (t, J=7.6 Hz, 6H),
1.28–1.40 (m, 4H), 1.44 (m, 2H), 1.92
(m, 2H), 2.32 (t, J=6.2 Hz, 2H), 2.38
(s, 6H), 2.44 (t, J=6.3 Hz, 4H), 2.49
Scheme 2. Reaction sequence for the synthesis of D.
(q, J=7.6 Hz, 4H), 2.82 (t, J=6.3 Hz,
4H), 4.01 (t, J=7.3 Hz, 2H), 7.31 (d,
J=8.1 Hz, 4H), 7.35 (dd, J=9.2,
(13 mL) and the compound was extracted with ethyl acetate (5”
200 mL). The organic solution was successively dried over anhydrous
Na2SO4 and vacuum evaporated to obtain a solid compound, which was
further dried under vacuum at room temperature in the presence of
KOH. Yield 11.6 g (37%); 1H NMR (CDCl3): d=1.38 (s, 9H), 1.42–1.50
(m, 8H), 2.32 (s, 2H), 2.65 (m, 2H), 3.04 (m, 2H), 4.67 ppm (s, 1H); 13C
(CDCl3): d=26.4, 26.5, 28.4, 30.0, 33.3, 40.4, 41.9, 78.9, 155.9 ppm; ele-
mental analysis calcd (%) for C11H24N2O2: C 61.08, H 11.18, N 12.95;
found: C 61.02, H 11.23, N 12.89.
2.1 Hz, 2H), 7.68 (d, J=8.1 Hz, 4H), 8.26 (d, J=9.2 Hz, 2H), 8.27 ppm
(d, J=2.1 Hz, 2H); 13C NMR (CD3OD): d=9.8, 21.5, 27.5, 27.6, 27.8,
30.7, 31.3, 42.0, 50.1, 54.7, 55.0, 106.1, 109.7, 117.5, 127.5, 128.0, 130.7,
138.6, 142.8, 144.6, 145.7, 156.9, 176.0 ppm; elemental analysis calcd (%)
for C43H55N7O6S2: C 62.22, H 6.68, N 11.81; found: C 62.11, H 6.71, N
11.76.
Synthesis of D·4HBr·EtOH: A solution containing 6 (1.90 g, 0.0023 mol)
and phenol (10.8 g, 0.11 mol) in 33% HBr/CH3COOH (370 mL) was
maintained at 908C under stirring for 24 h. The resulting suspension was
then cooled to room temperature and CH2Cl2 (350 mL) was added to
complete the precipitation of the light-orange compound, which was fil-
tered, washed with CH2Cl2, recrystallised from an EtOH/H2O (3:1 v/v)
mixture and dried in vacuum at 308 K in the presence of KOH to obtain
Synthesis of tert-butyl{6-
ACHTREUNG
amino]hexyl}carbamate (3):
A
0.056 mol) in dry acetonitrile (140 mL) was added to a refluxing solution
of 2 (5.0 g, 0.023 mol) in dry acetonitrile (140 mL) over a period of 3 h.
The solution was refluxed for further 3 h, then the solvent was removed
under reduced pressure to obtain a pale-yellow oil, which was successive-
ly subjected to chromatography on silica gel (CH2Cl2/MeOH 100:3) mix-
ture. The eluted fractions were combined and evaporated to dryness
under reduced pressure. Yield: 7.35 g, 85%; 1H NMR (CDCl3): d=1.09
(m, 2H), 1.20 (m, 4H), 1.39(m, 2H), 1.42 (s, 9H), 2.17 (m, 2H), 2.40 (s,
6H), 2.42 (m, 4H), 2.86 (m, 4H), 3.04 (m, 2H), 4.63 (s, 1H), 5.61 (s, 2H),
7.28 (d, J=8.0 Hz, 4H), 7.76 ppm (d, J=8.0 Hz, 4H); 13C (CDCl3): d=
21.6, 26.0, 26.5, 26.9, 28.5, 29.9, 40.6, 53.1, 53.4, 79.0, 127.0, 129.6, 136.7,
143.1, 155.8 ppm; elemental analysis calcd (%) for C29H46N4O6S2: C
57.02, H 7.59, N 9.17; found: C 57.07, H 7.53, N 9.24.
1
D·4HBr·EtOH. Yield 0.80 g, 45%; H NMR (D2O, pH 3.8): d=1.12–1.68
(m, 8H), 2.52 (m, 2H), 2.80 (m, 4H), 3.11 (m, 4H), 3.19(m, 2H), 5.96
(d, J=2.2 Hz, 2H), 6.45 (dd, J=9.5, 2.2 Hz, 2H), 7.28 ppm (d, J=9.5 Hz,
2H); 13C NMR (D2O, pH 3.8): d=25.2, 26.7, 26.9, 30.1, 37.5, 48.8, 50.8,
52.8, 97.1, 104.3, 114.3, 141.3, 153.0 ppm; MS: m/z: 410.36 [M+H]+; UV/
Vis (H2O, pH 2.1, 298 K): lmax (e)=383 (25500), 275 nm
(50200 molꢀ1 dm3 cmꢀ1); elemental analysis calcd (%) for C25H45N7Br4O:
C 38.53, H 5.82, N 12.58, Br 45.06; found: C 38.61, H 5.77, N 12.73, Br
44.95 (the bromide content was evaluated by potentiometric titrations of
the compound with AgNO3).
Synthesis of N,N’-{[(6-aminohexyl)imino]diethane-2,1-diyl}bis(4-methyl-
benzenesulfonamide) (4): CF3COOH (150 mL) was added dropwise to a
solution of 3 (7.35 g, 0.012 mol) in CH2Cl2 (120 mL) at 08C (water/ice
bath). The resulting solution was maintained at room temperature for
24 h, then cooled in a water/ice bath and alkalinised by the addition of
2m NaOH (660 mL). The organic phase was separated, and the aqueous
phase was extracted with CH2Cl2 (3”200 mL). The organic solutions
were combined and dried over anhydrous Na2SO4. The solvent was then
eliminated by evaporation under reduced pressure to obtain compound
Potentiometric measurements: All pH-metric measurements (pH=
ꢀlog[H+]) that were employed for the determination of equilibrium con-
stants were carried out in 0.1m NMe4Cl solutions at 298.1ꢁ0.1 K, by
using equipment and methodology already described.[57] The combined
Hamilton glass electrode (LIQ-GLASS 238000/08) was calibrated as a
hydrogen concentration probe by titrating known amounts of HCl with
CO2-free NaOH solutions and by determining the equivalent point by
Granꢁs method[58] which allows one to determine the standard potential,
Eo, and the ionic product of water (pKw =13.83(1) at 298.1 K in 0.1m
NMe4Cl). At least three measurements were performed for each system
in the pH ranges 2.5–10.5. In ligand protonation experiments, the concen-
tration of ligand CD was 5”10ꢀ4 m. In complexation experiments the con-
1
4, as a sticky yellow oil, in quantitative yield. H NMR (CDCl3): d=1.02–
1.34 (m, 6H), 1.39(m, 2H), 2.18 (m, 2H), 2.38 (m, 4H), 2.40 (s, 6H),
2.68 (m, 2H), 2.86 (m, 4H), 3.86 (s, 4H), 7.29(d, J=8.0 Hz, 4H),
194
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 184 – 196

DNA Binding by a Metallointercalator
FULL PAPER
centration of D was varied in the range 1”10ꢀ4–1”10ꢀ3 m, while the con-
centration of ZnII was CD ꢄCZnII ꢄ2CD. The computer program HYPER-
QUAD[59] was used to calculate the equilibrium constants from the elec-
tromotive force data.
chosen according to the intercalative and non-intercalative binding
modes.
Upon completion of each docking calculation, twenty poses per ligand
were saved. The overall sixty conformers that were obtained for H3D3+
were clustered on the basis of the position that was assumed with respect
to the DNA double helix and were ranked according to the glide score.
The top-ranked conformer per cluster was selected. The DNA–ligand
Spectral measurements: Absorption spectra were recorded on a Perkin–
Elmer Lambda 35 spectrophotometer and fluorescence emission spectra
on a Perkin–Elmer LS 55 spectrofluorimeter at 298.1ꢁ0.1 K. The titra-
tion experiments for ligand protonation and ZnII-complexation studies
were performed in 0.1m NMe4Cl solution. Spectrophotometric titrations
of the dye alone and its ZnII complex with calf thymus DNA were carried
out at I=0.11m (NaCl) and l=390 nm, by adding increasing amounts of
DNA solution directly into the cell containing the dye (CD =1”10ꢀ5 m) or
its ZnII-complex solution (C[ZnD] =1”10ꢀ5 m).
complexes were submitted to
a
minimisation with Impact v. 4.0,[63]
OPLS2005 force-field-based conjugate gradient minimisation routine
(atomic charges from the FF, maximum number of minimisation cycle:
5000; energy change criterion: 0.01 kcalmolꢀ1
Š
ꢀ1; gradient criterion:
0.01). Continuum representation of the solvent have been used by means
of the Poisson–Boltzmann solver equation.[64]
Fluorescence titrations of DNA samples were performed at lexc =385 nm
and lem =480 or 500 nm by using the procedure above described.
Circular dichroism spectra were recorded on a Jasco J710 spectropolarim-
eter (Easton, MD, USA).
Acknowledgement
Kinetic measurements: The kinetic measurements were carried out on a
Biologic SFM 300 stopped-flow mixing unit coupled to a spectrophoto-
metric line by two optical guides. A blue laser diode (l=405 nm, 1 mW)
was used as a light source. The radiation was passed through a Bausch
and Lomb 338875 high-intensity monochromator and then split into two
beams. The reference beam was sent directly to a 1P28 photomultiplier.
The measuring beam was sent through an optical quartz guide to the ob-
servation chamber and then, through a second optical guide, to the meas-
uring photomultiplier, which was also IP28. The output from the two
photomultipliers was balanced before each shot. The stopped-flow appa-
ratus is able to measure absorbance and/or fluorescence changes, but
only the fluorescence detection mode was employed. The acquisition
system keeps a record of a number of data points ranging from 10–8000
with sampling intervals ranging from 50 ms–10 s. Each experiment was re-
peated at least ten times, and the relaxation curves obtained were aver-
aged by an accumulation procedure. All measurements were carried out
under pseudo-first-order conditions.
The authors wish to thank Prof. S. Yarmoluk for the kind gift of a
Cyan40 sample.
[1] S. Neidle, Prog. Med. Chem. 1979, 16, 151–221.
[2] H. M. Berman, P. R. Yuong, Annu. Rev. Biophys. Bioeng. 1981, 10,
87–114.
[3] W. P. Wilson, R. L. Jones, Adv. Pharmacol. Chemother. 1981, 18,
177–222.
[4] R. Zittoun, Eur. J. Cancer Clin. Oncol. 1985, 21, 649–653.
[5] A. Albert, The Acridines: Their Preparation, Physical, Chemical,
and Biological Properties and Uses, Arnold, London, 1966.
[6] L. S. Lerman, J. Mol. Biol. 1961, 3, 18–30.
[7] A. Aggarwal, S. A. Islam, R. Kuroda, S. Needle, Biopolymers 1984,
23, 1025–1041.
[8] A. Adams, Curr. Med. Chem. 2002, 9, 1667–1675.
[9] L. Malinina, M. M. Soler-Lòpez, J. Aymamì, J.A. Subirana, Bio-
chemistry 2002, 41, 9341–9348.
[10] L. P. G. Wakelin, M. J. Waring, J. Mol. Biol. 1980, 144, 183–214.
[11] S. Lippard, Acc. Chem. Res. 1978, 11, 211–217.
[12] C. Hiort, P. Lincoln, B. Norden, J. Am. Chem. Soc. 1993, 115, 3448–
3454.
[13] S. Arounaguiri, D. Easwaramoorthy, A. Ashokkumar, A. Dattagup-
ta, G. M. Bhaskar, Proc. Indian Acad. Sci. Chem. Sci. 2000, 112, 1–
17.
[14] A. H. Krotz, B. P. Hudson, J. K. Barton, J. Am. Chem. Soc. 1993,
115, 12577–12578; R. H. Terbrueggen, J. K. Barton, Biochemistry
1995, 34, 8227–8234.
Unwinding measurements: Supercoiled pBR322 DNA (0.15 mg) was incu-
bated with 1 U of topoisomerase I (Invitrogen) in Tris HCl (50 mm),
pH 7.5, KCl (50 mm), MgCl2 (10 mm), DTT (0.5 mm), EDTA (100 mm),
BSA (30 mgmLꢀ1) in the presence/absence of increasing concentrations
of D or [ZnD]. After overnight incubation at 310 K, NaCl was added to
the samples to a final concentration of 0.15m and, the samples were ex-
tracted with an equal volume of 1:1 PhOH/CHCl3 that was buffered to
pH 7.5 with Tris HCl (50 mm), to eliminate both the dye and protein. The
aqueous phase was loaded on 1% agarose gel, and run in TAE (40 mm
Tris, 18 mm acetic acid, 1 mm EDTA). The reaction products were visual-
ised by ethidium bromide staining.
[15] K. E. Erkkila, D. T. Odom, J. K. Barton, Chem. Rev. 1999, 99, 2777–
2795.
Molecularmodelling : Computations were performed on AMD 1 GHz
and Intel P4 3 GHz computers running Linux.
[16] E. J. C. Olson, D. Hu, A. Hoermann, A. M. Jonkman, M. R. Arkin,
E. D. A Stemp, J. K. Barton, P. F. Barbara, J. Am. Chem. Soc. 1997,
119, 11458–11467.
[17] A. C. Lim, J. K. Barton, Bioorg. Med. Chem. 1997, 5, 1131–1136.
[18] B. A. Jackson, J.K. Barton, J. Am. Chem. Soc. 1997, 119, 12986–
12987; B. A. Jackson, V. Y. Alekseyev, J.K. Barton, Biochemistry
1999, 38, 4655–4662.
The conformational search was performed by using the simulated anneal-
ing (SA) procedure contained in MacroModel version 8.5,[60] by using the
MMFF force field (FF) and the atomic partial charges from FF. The set-
tings of the molecular dynamic simulation were as follow: running tem-
perature: T=400 K; heating, running and cooling time: 5 ps; time step:
1.5 fs. 100 Energy-minimised conformers were obtained and collected in
cluster by using an “in house”-developed routine that was written in
Visual Basic programming language version 6.0 (Microsoft Corporation).
The routine groups molecules on the basis of the root mean square
(RMS) value of the distance between atoms (RMSꢄ0.9Š).
DNA decamers in the B-form were built by using the Builder module in
Maestro version 7.5[61] for all the following sequences: ATATATATAT,
GCGCGCGCGC, ATCGCGCGAT, CGCGATATCG, CGATCGATCG.
Intercalation sites of CG and/or AT types were conveniently built in the
centre of the decamers for each DNA sequence.
[19] L. G. Marzilli, S. Ano, F. P. Intini, G. Natile, J. Am. Chem. Soc. 1999,
121, 9133–9142.
[20] E. A. Boudreaux, Int. J. Quantum Chem. 2001, 83, 255–258.
[21] B. E. Bowler, K. J. Ahmed, W. I. Sundquist, L. S. Hollis, E. E.
Whang, S. J. Lippard, J. Am. Chem. Soc. 1989, 111, 1299–1306.
[22] C. Ciatto, M. L. D’Amico, G. Natile, F. Secco, M. Venturini, Bio-
phys. J. 1999, 77, 2717–2724.
[23] M. E. Lamm, D. M. Neville, Jr., J. Phys. Chem. 1965, 69, 3872–3877.
[24] R. W. Chambers, T. Kajiwara, D. R. Kearns, J. Phys. Chem. 1974, 78,
380–387.
[25] H. Porumb, Prog. Biophys. Mol. Biol. 1978, 34, 175–195.
[26] D. J. Blears, S. S. Danyluk, J. Am. Chem. Soc. 1967, 89, 21–26.
[27] R. Luchowski, S. Krawczyk, Chem. Phys. 2003, 293, 155–166.
[28] A. N. Glazer, Proc. Natl. Acad. Sci. USA 1965, 54, 171–176.
Docking calculations were performed by using the Glide program.[62] De-
fault input parameters were used in all computations (no scaling factor
for the vdW radii of nonpolar DNA atoms, 0.8 scaling factor for nonpolar
ligand atoms). The grids for the docking (20 Š edges for the enclosing
box) were centred on the centroid of the appropriate residues, and were
Chem. Eur. J. 2008, 14, 184 – 196
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
195

A. Bianchi, F. Secco et al.
[29] R. M. Smith, A. E. Martell, NIST Stability Constants Database, ver-
sion 4.0, National Institute of Standards and Technology, Washing-
ton, DC, 1997.
[30] B. Adcock, in Acridines, 2nd ed. (Ed.: R. M. Acheson) Interscience,
New York, 1973, pp. 109–137.
[46] H. Strehlow, Rapid Reaction in Solution, Wiley-VCH, Weinheim,
1992, p. 16.
[47] H. Baruah, U. Bierbach, Nucleic Acids Res. 2003, 31, 4138–4146.
[48] A. Adams, J. M. Guss, C. A. Collyer, W. A. Denny, L. P. G. Wakelin,
Biochemistry 1999, 38, 9221–9223.
[49] C. A. Frederick, L. D. Williams, G. Ughetto, G. A. Van der Marel,
J. H. Van Boom, A. Rich, A. H. J. Wang, Biochemistry 1990, 29,
2538–2549.
[50] N. J. Pritchard, A. Blake, A. R. Peacocke, Nature 1966, 212, 1360–
1361.
[51] A. Blake, A. R. Peacocke, Biopolymers, 1968, 6, 1225–1253.
[52] “Techniques and Applications of Fast Reactions in Solution”: T. Ya-
nusaga, H. Ushio, in , NATO ASI Ser. Ser. C1979, 50, 389–397.
[53] M. Kamiya, Biochim. Biophys. Acta 1978, 517, 527–530.
[54] S. S. Lukashov, M. Y. Losytskyy, Y. L. Slominskii, S. M. Yarmoluk,
Biopolym. Cell. 2001, 17, 169–177.
[55] A.K. Covington, M. Paabo, R. A. Robinson, R. G. Bates, Anal.
Chem. 1968, 40, 700–706.
[56] S. A. Gamage, N. Tepsiri, P. Wilairat, S. J. Wojcik, D. P. Figgitt, R. K
Ralph, W. A. Denny, J. Med. Chem. 1994, 37, 1486–1494.
[57] A. Bianchi, L. Bologni, P. Dapporto, M. Micheloni, P. Paoletti,
Inorg. Chem. 1984, 23, 1201–1205.
[31] T. Biver, F. Secco, M. R. Tin›, M. Venturini, Arch. Biochem. Bio-
phys. 2003, 418, 63–70.
[32] H. J. Li, D. M. Crothers, J. Mol. Biol. 1969, 39, 461–477.
[33] M. Dourlent, J. F. Hogrel, Biochemistry 1976, 15, 430–436.
[34] a) W. C. Tse, D. L. Boger, Acc. Chem. Res. 2004, 37, 61–69 ; b) T.
Biver, D. Lombardi, F. Secco, M. R. Tin›, M. Venturini, A. Bencini,
A. Bianchi, B. Valtancoli, Dalton Trans. 2006, 10, 1524–1533;
c) T. C. Jenkins, Methods in Molecular Biology, Humana Press,
Totowa, 1997, pp. 195–218.
[35] D. G. Dalgleish, H. Fujita, A. R. Peacocke, Biopolymers, 1969, 8,
633–645.
[36] S. M. Zeman, K. M. Depew, S. J. Danishefsky, D. M. Crothers, Proc.
Natl. Acad. Sci. USA 1998 95, 4327–4332.
[37] ConQuest Version 1.8, Cambridge Crystallographic Data Centre, 12
Union Road CB 1EZ, UK.
[38] VISTA version 2.1, Cambridge Crystallographic Data Centre, 12
Union Road CB 1EZ, UK.
[39] a) CSD version 5.27 (November 2005); b) F. H. Allen, J. E. Davies,
J. J. Galloy, O. Johnson, O. Kennard, C. F. Macrae, E. M. Mitchell,
G. F. Mitchell, J. M. Smith, D. G. Watson, J. Chem. Inf. Comput. Sci.
1991, 31, 187–204.
[40] J. N. Lisgarten, M. Coll, J. Portugal, C. W. Wright, J. Aymami, Nat.
Struct. Biol. 2002, 9, 57–60.
[58] a) G. Gran, Analyst (London) 1952, 77, 661–671; b) F. J. Rossotti,
H. Rossotti, J. Chem. Educ. 1965, 42, 375–378.
[59] P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43, 1739–1753.
[60] a) F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M.
Lipton, C. Caufield, G. Chang, T. Hendrickson, W. C. Still, J.
Comput. Chem. 1990, 11, 440–467; b) L. L. C. Schrçdinger, New
York http://www.schrodinger.com.
[41] a) S. Neidle, T. A. Jones, Nature 1975, 253, 284–285; b) S. K. Oben-
dorf, H. L. Carrell, J. P. Glusker, Acta Crystallogr. Sect. B 1974, 30,
1408–1411.
[42] a) E. R. Gordon, R. B. Walsh, W. T. Pennington, T. W. Hanks, J.
Chem. Crystallogr. 2003, 33, 385–390; b) C. A. Mattia, L. Mazzarel-
la, V. Vitagliano, R. Puliti, J. Crystallogr. Spectrosc. Res. 1984, 14,
71–87.
[61] Maestro v 7.5, L. L. C. Schrçdinger, New York http://www.schrodin-
ger.com.
[62] a) R. A. Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. Klicic,
D. T. Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry,
D. E. Shaw, P. Francis, P. S. Shenkin, J. Med. Chem. 2004, 47, 1739–
1749; b) L. L. C. Schrçdinger, New York http://www.schrodinger.-
com.
[43] a) R. Puliti, C. A. Mattia, L. Mazzarella, Acta Crystallogr., Sect. C
2002, 58, 582–584; b) R. Puliti, C. A. Mattia, Acta Crystallogr. Sect.
A 2001, 57, 1338–1340.
[63] Impact v 4.0, L. L. C. Schrçdinger, New York http://www.schrodin-
ger.com.
[64] M. C. Cortis, R. A. Friesner, J. Comput. Chem. 1997, 18, 1591–1608.
[44] H. Li, X. Peng, P. Leonard, F. Seela, Bioorg. Med. Chem. 2006, 14,
4089–4100.
[45] T. Biver, A. De Biasi, F. Secco, M. Venturini, S. Yarmoluk, Biophys.
J. 2005, 89, 374–383.
Received: December 21, 2006
Revised: June 13, 2007
Published online: September 13, 2007
196
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 184 – 196